The problem with statistics and numbers is this: you need to know where they came from and what they are actually telling you or they are worth nothing.
I like to back up the claims I make in my blog posts with numbers. Statistics are a great way to add factual weight to what I am saying, and most other content creators do the same. But often the number used to build credibility (or bragging rights) simply doesn’t tell the story it is claimed to tell.
The AI Statistic Problem
I ran into this myself while researching for a blog post. To speed things up, I asked ChatGPT, Claude, and Gemini for statistics to support my claims. They gave me a list of numbers and statements to choose from – they sounded great.
When I asked them to provide a source for one I liked, the answer was something like this: “The statistic is frequently cited in marketing discussions. However, it does not specify the original study or data from which this statistic is derived.”
Nice numbers, no source = not worth anything.
I am familiar with this problem. And it isn’t just an AI problem.
When I tried to search for an original source for a statistic that I found, I clicked from blog post to article to blog post – all citing another blog post as source. But there was never a mention of WHO made the study and HOW!
And the problem with statistics is not just a lack of a source:
The Interpretation Problem
You also need to think about interpretation of numbers:
A while ago there used to be yearly comparisons of universities, mine ended up fairly bad for IT which was my second subject so I looked into it. They had compared the universities by the number of students finishing with a degree
My university had just opened the IT department so nobody finished with a degree – yet.
The Marketing Spin Problem
And it is not just statistics. People like to play around with numbers to impress, but when you look a little closer there is often not much impressive facts in it:
When I ran my startup, news outlets liked to cover startups for big growth: The article then mentioned something like ‘They doubled their customers in a year.’
Sounds great, but when you start with 1 customer, 2 customers a second year is not that impressive.
A big-name marketer used to write articles about how Facebook was so much better than Twitter and mentioned how much traffic and signups he got from Facebook. He somehow ‘forgot’ to mention that he ran ad-campaigns on Facebook for several hundred thousand dollars.
The Correlation Problem
My brother told me about a study in Scandinavia that found that people who get flu vaccinations tend to have fewer heart problems. He took this to mean that the vaccine itself protects against heart failure.
I wasn’t convinced.
I wasn’t questioning the study. I was questioning the conclusion.
There is a connection between influenza and cardiovascular complications. So perhaps the relevant factor isn’t some mysterious heart-protective effect of the vaccine at all. Perhaps part of the effect comes from the fact that vaccinated people get influenza less often.
Either way, the number alone doesn’t tell us which explanation is correct.
Back to marketing and business: All these ‘I made this huge impressive amount of money in 4 hours’ statements hide the fact that they built an audience and a brand or reputation for months or even years before this happened.
A number doesn’t exist in isolation.
Behind every statistic is a method, a sample, a definition, a time period and a context. And once the number is published, there is another layer: the story someone tells about it.
That is where things get interesting.
The number is one thing. The meaning we attach to it is another.
I have a PhD in math, and this still gets me:
Sometimes the number is wrong. Sometimes the number is perfectly correct but measures something different from what you think. Sometimes the number is technically correct but missing the context that would change your interpretation. And sometimes the conclusion simply doesn’t follow from the number.
So which numbers can you trust?
I trust numbers much more when I can trace them back to:
- the original research — ideally the actual study rather than an article talking about the study
- a transparent report or dataset — where you can see who collected the data and how
- a documented case study — especially when the company explains what happened, over what period, and gives enough context to understand the numbers. A case study can tell you what happened. It doesn’t automatically tell you why it happened or whether you can reproduce it.
- official statistics — when the organization collecting the data is appropriate for the question
- multiple independent sources that arrive at similar findings
But even then, I don’t automatically believe the conclusion.
I want to know what was actually measured, how the number was calculated, how large the underlying sample was, what it was compared with, and whether the conclusion drawn from it actually follows.
When you see an impressive number, ask:
Where does it come from?
Can I find the original research, report, dataset or case study?
What exactly does it measure?
Is it actually measuring what the writer claims it measures?
Compared to what?
A 100% increase means very different things depending on the starting point.
What’s missing?
Costs? Time? Sample size? Advertising spend? Previous work?
Does the conclusion follow?
Or has somebody turned a correlation into a causal claim?
So next time you see a great statistic or an impressive number, don’t just ask whether it is true.
Ask what story the number is being used to tell you.
Then ask whether the number actually supports that story.
Because numbers can make almost any claim sound more credible.
That doesn’t mean the claim is true.
And sometimes, the most deceptive number is the one that is completely honest.
If this kind of practical, no-fluff thinking is useful to you, I share one content creation or optimization tip a day in my free 20 Days of Content Brilliance sequence.
